Scribd
Upload
6 views1 page

Tutorial 3

This document is a tutorial on Numerical Linear Algebra, focusing on problems related to stability in algorithms. It includes axioms and tasks that require proving the stability of various algorithms and functions involving complex numbers. The tutorial emphasizes backward stability and provides specific problems for students to solve under the guidance of an instructor and tutor.

Uploaded by

ch23bt004
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views1 page

Tutorial 3

This document is a tutorial on Numerical Linear Algebra, focusing on problems related to stability in algorithms. It includes axioms and tasks that require proving the stability of various algorithms and functions involving complex numbers. The tutorial emphasizes backward stability and provides specific problems for students to solve under the guidance of an instructor and tutor.

Uploaded by

ch23bt004
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

Introduction to Numerical Linear Algebra

Tutorial - 3
Instructor: Dr. Rajshekar V Bhat
Tutor: Periyasamy M

Tutorial problems
1. Show that (1 + O(ϵmachine ))(1 + O(ϵmachine )) = 1 + O(ϵmachine ). The precise meaning of
this statement is that if f is a function satisfying f (ϵmachine ) = (1 + O(ϵmachine ))(1 +
O(ϵmachine )) as ϵmachine → 0, then f also satisfies f (ϵmachine ) = 1 + O(ϵmachine ) as
ϵmachine → 0.
2. Show that (1 + O(ϵmachine ))−1 = 1 + O(ϵmachine ).

Axioms
We need the following axioms to solve the upcoming problems.
(a) For all x ∈ R(C as well), there exists ϵ with |ϵ| ≤ ϵmachine such that f l(x) = x(1+ϵ).

(b) For all x,y ∈ F, there exists |ϵ| ≤ ϵmachine such that x ○
∗ y = (x ∗ y)(1 + ϵ).
3. Let f : C2 → C be the problem defined by f (x1 , x2 ) = x1 − x2 and the algorithm is
given by f˜(x1 , x2 ) = f l(x1 ) ⊖ f l(x2 ). Prove that the algorithm f˜ is backward stable.
4. Prove that the standard inner product on C2 is backward stable (if x,y∈ C2 then the
standard inner product is defined by α(x, y) = x∗ y).
5. Let f : C → C be the problem defined by f (x) = x + 1 and the algorithm is given by
f˜(x) = f l(x) ⊕ 1. Prove that the algorithm f˜ is stable, but not backward stable.
6. State whether the following algorithms are backward stable, stable but not backward
stable or unstable, and prove it or at least give a reasonably convincing argument.
(i) Data: x ∈ C. Solution: 2x, computed as x ⊕ x.
(ii) Data: x ∈ C. Solution: x2 , computed as x ⊗ x.
(iii) Data: x ∈ C \ {0}. Solution: 1, computed as x ○
÷ x.
(iv) Data: x ∈ C. Solution: 0, computed as x ⊖ x.

You might also like